|
|
|
|||||||
|
|||||||||
(Received for publication, September 27, 1994) From the
It has previously been shown that very low steady state levels
of Type I collagen, the most abundant collagen found in
vertebrates, is a heterotrimer consisting of two In the case of the In the
case of the In the present study, the process by which the
All of the single-stranded probes were made
by subcloning in the vectors M13mp18 and M13mp19 as follows: the intron
1 probes (Int-mp18 and Int-mp19) contained the
Figure 1:
Nuclear run-on
assays using M13 single-stranded probes. A, chondrocytes, and B, BrdU-treated chondrocytes. 4 µg of each single-stranded
probe was used per slot. For the
Figure 3:
Map of portions of the
Figure 2:
RNase protection analyses. Strand-specific
riboprobes from the region of the
Fig. 2B shows the same experiment using
the corresponding riboprobe of antisense sequence, in order to detect
sense strand transcripts. As found previously(8) , very low
steady state levels of In addition to these RNase
protection analyses, the same analyses were also carried out using
riboprobes from the additional regions of the
On the other hand,
when antisense riboprobes were used, These results
are in agreement with the nuclear run-on assays, indicating that
antisense RNA is transcribed across the
Figure 4:
Control RNase protection experiments. A
riboprobe of sense strand sequence from the region of the
Since both of these RNase protection
experiments were performed using the same stringent hybridization
conditions as described for Fig. 2above, the possibility of
cross-hybridization was eliminated, and the hybridizing chondrocyte
RNAs must be
Figure 5:
Northern blot analyses. Strand-specific
probes from the 5` end (467-nt intron 1-exon 2 sequence) and 3` end
(345-nt carboxyl-terminal sequence) of the
When probes of antisense sequence were used (Fig. 5B), hybridization to
Figure 6:
RNase
protection analyses showing RNA decay. A, decay of the
Figure 7:
RNase protection analyses to test for the
accumulation of
Figure 8:
Genomic Southern blot analysis. Each lane contains 12 µg of chick liver DNA that had been
digested with either BamHI (B), EcoRI (E), HindIII (H), or StuI (S). A 467-nt probe containing the intron 1-exon 2 region was
used. The BamHI 0.63 kb and StuI 5.1 kb bands are
indicated.
Nuclear run-on experiments indicate that at the same
time that antisense transcription is occurring in chondrocytes, there
is a down-regulation of
Similar chondrocyte-specific
transcripts have also been detected in a previous study upon long
exposures of autoradiograms(3) , both with 5` and 3` end
double-stranded probes. Since, in the present study, these transcripts
are not detected with a probe specific for the sense mRNA, the previous
suggestion (3) that these are unspliced The reason for the existence of
such large transcripts in chondrocytes is unclear, but, while there may
be open reading frames in these transcripts, it has yet to be
determined if they are associated with the translational machinery of
the cell, even though preliminary data (not shown) indicate that they
are polyadenylated and are transported to the cytoplasm. It is also
possible that these large transcripts may fall into the class of
regulatory RNAs that are increasingly appearing in the
literature(28, 29) , perhaps playing a role in the
down-regulation of the Since the antisense transcripts have a
high transcription rate and yet are only accumulated at moderate levels
in chondrocytes, their stability was expected to be lower than that of
the type II collagen mRNA or the corresponding
One intriguing possibility is that the antisense
transcripts may regulate expression of the The fact that antisense transcription
continues through at least some of the promoter region (to position
-221, which includes the TATA box and an inverted CCAAT box) also
raises the possibility that antisense transcription can interfere with
initiation of It is also
possible, despite the low steady state level of If the antisense RNA does indeed
play a role in the down-regulation of the Whether or not the
chondrocyte-specific antisense transcripts play a functional role in
the down-regulation of the
Volume 270,
Number 7,
Issue of February 17, 1995 pp. 3400-3408
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
1(I)
Collagen Gene (*)
ABSTRACT
INTRODUCTION
MATERIALS AND METHODS
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES
1(I) collagen mRNA are present in chick embryo sternal
chondrocytes (Askew, G. R., Wang, S., and Lukens, L. N.(1991) J.
Biol. Chem. 266, 16834-16841), yet nuclear run-on
experiments with double-stranded cDNA probes indicated a high
transcription rate at this locus. These findings were investigated in
this study using single-stranded probes, where nuclear run-on
experiments showed that antisense transcription of the 1(I)
collagen gene was occurring in chondrocytes, while sense strand
transcription was down-regulated. Treatment of these chondrocytes with
5-bromo-2`-deoxyuridine (BrdU), which causes the cells to resemble
their mesenchymal precursors, resulted in an antiparallel situation,
where antisense transcription was lost, and instead, sense strand
transcription was acquired, suggesting that the reverse switch from
sense to antisense transcription occurs during chondrogenesis. Very
large (>10 kilobases) and heterogeneous antisense transcripts of
moderate stability were shown to span both ends of the gene in
chondrocytes, while their absence was shown in BrdU-treated
chondrocytes, chick embryo fibroblasts, and a variety of other tissues.
The function of these antisense transcripts is so far unknown, but
their unusual chondrocyte-specific appearance, concurrent with little
or no sense strand transcription, suggests a possible functional role
in the down-regulation of the 1(I) collagen gene.
1 chains and one
2 chain and is an important component of skin, bones, ligaments,
and tendons. Its expression can also be seen in cultured fibroblasts
and in fibroblast-like cells, such as chick embryo sternal chondrocytes
that have been treated with 5-bromo-2`-deoxyuridine (BrdU), (
)which induces chondrocytes to resemble the mesenchymal
cells from which they develop, in an apparent reversal of their
differentiation process(1, 2) . A similar type of
response can also be seen with chondrocytes that have been grown in the
presence of phorbol 12-myristate 13-acetate(3, 4) , or
with chondrocytes that have been transformed with Rous sarcoma
virus(5, 6, 7) . Untreated chick embryo
sternal chondrocytes do not produce type I collagen, but instead make
large amounts of type II collagen, a homotrimer consisting of three
1(II) chains and confined, almost exclusively, to hyaline
cartilage. During treatment of chondrocytes with BrdU over a period of
8 days, the 1(II) gene is down-regulated at the transcriptional
level(8) , while at the same time there is an up-regulation of
the two type I collagen genes.2(I) collagen
gene, this up-regulation is accomplished by a switch in promoter site
usage. The transcript found in chondrocytes, at approximately one-third
the level found in BrdU-treated chondrocytes, is shorter at the 5` end,
with the transcription start site occurring at a chondrocyte-specific
exon, named exon A, located in the normal intron 2 of the 2(I)
collagen gene(9, 10, 11) . Exon A is spliced
to exon 3 of this gene, in such a way that no open reading frame coding
for collagen sequences is present, so no 2(I) polypeptide is
produced in chondrocytes. Chondrocyte-specific promoter activity has
been found in the region located upstream of exon A(12) , but
this activity is lost upon treatment of chondrocytes with BrdU, and
instead, the normal 2(I) collagen gene promoter is used, with
transcription beginning at exon 1 (8, 11) .1(I) collagen gene, the steady state level of the mRNA
is very low in chick embryo sternal chondrocytes, being less than 2% of
the level found in BrdU-treated chondrocytes, as measured by a
single-stranded riboprobe in RNase protection experiments(8) .
However, nuclear run-on experiments, using double-stranded cDNA probes,
showed that the transcription rate at the 1(I) collagen locus was
as high in chondrocytes as in BrdU-treated chondrocytes(8) . In
addition, measurement of the cytoplasmic decay rate of the low level of
mRNA found in chondrocytes showed that the mRNA half-life (
12 h)
is the same as that found in BrdU-treated chondrocytes(8) .
This led to the suggestion that the 1(I) collagen gene is
transcribed at a high level in chondrocytes but that most of the mRNA
is rapidly broken down in the nucleus, perhaps due to a block in
splicing and/or nuclear-cytoplasmic transport. This suggestion was
supported by another study(3) , where Northern blots, again
using double-stranded cDNA probes, showed the presence of very large
(>10 kb) multiple transcripts in chondrocytes, but not in phorbol
12-myristate 13-acetate-treated chondrocytes, which phenotypically
resemble BrdU-treated chondrocytes. This finding led these authors to
suggest that these large RNAs represented unprocessed 1(I)
collagen transcripts.1(I) collagen gene is down-regulated in chondrocytes was examined.
It was found with the use of single-stranded probes that this
down-regulation is apparently not due to rapid nuclear degradation, as
originally suspected, but instead, due to the presence of antisense
transcription throughout the 1(I) collagen gene in chondrocytes,
with a simultaneous down-regulation of sense strand transcription. This
antisense transcription is responsible for the high rate of
transcription found in the previous nuclear run-on
experiments(8) , and the very large transcripts found in
Northern blots (3) , since double-stranded probes were used in
these previous studies, and therefore, antisense transcripts were being
detected. This report also shows that induction of sense strand
transcription in chondrocytes, by BrdU treatment, results in loss of
antisense transcription. Since BrdU-treated cells resemble
prechondrogenic mesenchymal cells, it is likely that antisense
transcription is acquired during chondrogenesis. The antisense
transcripts are also shown to be accumulated specifically in
chondrocytes and, so far, not in any other cell type examined. Since
relatively few cell or stage type-specific endogenous antisense RNAs
have been described in eukaryotes to date, this report represents a new
example and raises the possibility that antisense transcription
down-regulates transcription of the 1(I) collagen gene in
chondrocytes.
Cell Culture
Chondrocytes were obtained from the
sterna of 14-day-chick embryos (from Spafas, Norwich, CT) by the
floater selection method(13) , after culturing them for 3 days
in Dulbecco's modified Eagle's medium supplemented with 10%
fetal calf serum (Life Technologies, Inc.). For BrdU treatment, these
chondrocytes were replated at 1.5 
10
cells/ml in
Dulbecco's modified Eagle's medium, 10% serum containing
6.5 10
M bromodeoxyuridine
(Boehringer Mannheim) and allowed to incubate for a further 8-10
days, with addition of 1 ml of fresh medium plus BrdU every 2 days. For
actinomycin D treatment, the chondrocytes were cultured for 3 days as
usual, and then actinomycin D (Sigma) was added directly to each plate
to give a final concentration of 5 µg/ml, after which the cells
were incubated for the required time interval. Chick embryo fibroblasts
were obtained from 10-day-old embryos and cultured as described
previously(11) , and calvaria were obtained from 14-day-old
embryos and cultured as described by Pawlowski (14) .Plasmids and M13 Clones
All of the 5` end genomic
subclones of the 1(I) collagen gene were derived from the clones
pRS4.6 and pRS500(15) , a gift from Louis Gerstenfeld. For use
as riboprobe templates, all sequences were inserted in either
pBluescript SK (pBt) or pBS (Stratagene). pBt-314 was made by inserting
the BamHI to AccI fragment (from position -221
in the promoter to 93 nt into exon 1) into the same sites in pBt;
pBt-Int contained the 550-nt HinfI (blunt-ended) to BamHI fragment of intron 1, inserted into the HincII
and BamHI sites of pBt; pBS-466 contained the BamHI
(in intron 1) to SalI (end of exon 2) fragment inserted in the BamHI and SalI sites of pBS; pBtRS500 contained the
428 nt SalI to EcoRI fragment, from the end of exon 2
to 83 nt into exon 5, inserted in the same sites of pBt.
pBSCg54-344 (1(I) collagen cDNA containing 345 nt of
carboxyl-terminal sequence beginning 21 nt after the triple helical
coding region) and pBSCg12-215 (type II collagen cDNA containing
215 nt of the 3`-untranslated region beginning 910 nt after the
termination codon) were made as described previously(8) . The
0.9-kb PstI insert of the 27 S cDNA clone p11D2 (8) was subcloned in pBt for use in nuclear run-on experiments
(pBt-27S), and a 180-nt PstI to BamHI
subfragment of this cDNA was inserted in pBt (pBt-27SBam) for synthesis
of the 27 S riboprobe.573-nt KpnI to BamHI insert (includes 23 nt of the pBt
polylinker sequence) of pBt-Int; the intron 1-exon 2 probes (466-mp18
and 466-mp19) contained the 467-nt BamHI to SalI
insert of pBS-466; the intron 2-exon 5 probes (500-mp18 and 500-mp19)
contained the 428-nt SalI to EcoRI insert of
pBtRS500; the carboxyl-terminal probes (344-mp18 and 344-mp19)
contained the 370-nt HindIII to SacI insert (includes
26 nt of the pBS polylinker sequence) of pBSCg54-344; the type II
collagen 3` end probes (215-mp18 and 215-mp19) contained the 264-nt HindIII to SacI insert (includes 49 nt of the pBS
polylinker sequence) of pBSCg12-215; and the fibroblast-specific
2(I) collagen 5` end probes (4432-mp18 and 4432-mp19)
were derived from the genomic clone pCg5.7(16) , where an
854-nt SmaI fragment (from the beginning of exon 1 to the
middle of intron 1) was first inserted in the HincII site of
pBt, and from this a 466-nt fragment (starting at the KpnI
site in the pBt polylinker sequence, spanning the exon 1 sequence, and
ending at the EcoNI site in intron 1) was inserted in the KpnI and HincII sites of the M13 vectors.Nuclear Run-on Assays
Nuclei were isolated and
run-on reactions were performed as described previously(8) ,
except that heparin was not used in the reactions, and
Inhibit-ACE
(5 Prime
3 Prime, Boulder, CO), 7.5
units/reaction, was used as an RNase inhibitor instead of RNasin.
Hybridizations of the 
P-labeled run-on transcripts were as
described previously (8) , except that all probes were
single-stranded in M13 vectors (except for the 27 S cDNA), and were
slot-blotted in triplicate on nitrocellulose membranes (Schleicher and
Schuell). Additional washing steps were also carried out in 0.2
SSC, 0.5% SDS, 1 30 min, 65 °C, and 0.2 SSC, 0.1%
SDS, 1 30 min, 65 °C.RNA and Genomic DNA Isolation
In most cases, total
cellular RNA was isolated as described previously(17) , or
alternatively, in the case of RNA extraction from various tissues, by
the Promega procedure (18) , where freshly dissected tissues
from 18-day-old embryos were first minced in chilled denaturing
solution in a Dounce homogenizer. In the case of the blood cell RNA
preparation, whole blood cells were first spun down at 1,000 g, 4 °C, before disruption in denaturing solution. For the
actinomycin D chase experiment, drug-treated chondrocytes from five
100-mm tissue culture dishes were chilled on ice at the appropriate
time points, and total RNA was isolated as described
previously(17) . Genomic DNA was isolated from the livers of
14-day-old chick embryos by the procedure of Blin and
Stafford(19) , with addition of a CsCl gradient purification
step as described by Maniatis et al.(20) .RNase Protection Assays
Riboprobes were
synthesized as described previously(8, 11) , using
either T3 or T7 polymerases, depending on the orientation of the
template inserts in pBS or pBt, and whether sense or antisense sequence
riboprobes were required. RNase protection assays were carried out as
described(8, 11) , except that hybridizations for
1(I) collagen probes were performed at 68 °C (or 55 °C in
the case of the 27 S rRNA and type II collagen probes), and RNase
digestions were for 1 h at 30 °C using the recommended levels of
RNase A and RNase T1 (21) in order to ensure adequate digestion
of single-stranded RNAs. For the control experiments with RNase V1 and
RNase H (both from United States Biochemical Corp.), the RNase
protection assays were carried out as usual, but instead of dissolving
the final pellets in gel loading dye, they were treated as follows. For
the RNase V1 controls all pellets were dissolved in 200 µl of
1 RNase V1 buffer (20 mM Tris-HCl (pH 7.6), 200 mM NaCl, 10 mM MgCl
), and for the RNase H
controls in 200 µl of 1 RNase H buffer (20 mM Tris-HCl (pH 7.6), 100 mM KCl, 10 mM MgCl, 1 mM dithiothreitol). Enzyme was added
to the appropriate samples (30 units/ml for RNase V1 or 10 units/ml for
RNase H), and all samples were incubated for 30 min at 37 °C.
Following one phenol/chloroform and one chloroform extraction, samples
were ethanol precipitated, pellets were washed in 75% ethanol,
resuspended in 10 µl of gel loading dye, denatured, and run on 6%
sequencing gels as usual.Northern and Southern Blots
RNA samples were
separated on 0.8% agarose, 2.2 M formaldehyde gels according
to standard procedures(22) , and blots were carried out
according to instructions given with the Gene-Lite Chemiluminescent Detection kit (Bio-Rad), except that the
concentration of NaHPO
was 0.25 M in
the primary hybridization solution and 20 mM in the primary
hybridization wash solutions. The primary probes were cloned in M13
vectors as described above. For Southern blot analysis, genomic DNA was
digested overnight with the appropriate restriction endonuclease and
separated on a 0.8% agarose gel. Blotting was performed according to
the Gene-Lite kit, using the same hybridization and wash conditions as
for the Northern blots.
Antisense Transcription Occurs Across a Major Portion
of the
In order to
investigate the high level of transcription at the 1(I) Collagen Gene in Chondrocytes1(I) collagen
locus, nuclear run-on assays were performed with chondrocyte nuclei
using single-stranded probes in M13 vectors. Probes spanning four
different regions of the 1(I) collagen gene were used; three of
these were consecutive 5` end probes spanning from the HinfI
site in intron 1 to the EcoRI site 83 nt into exon 5, and one
was a 3` end probe containing 345 nt of the carboxyl-terminal sequence,
as described in Fig. 1and mapped in Fig. 3. In addition,
probes for the fibroblast-specific 5` end of the 2(I) collagen
gene and the 3`-untranslated region of the type II collagen gene were
used as controls for cell type specificity, and a cDNA probe for 27 S
ribosomal RNA was used as a positive control for all cell types. In
these nuclear run-on assays, all four of the 1(I) collagen probes
showed that predominantly antisense transcription was occurring across
the 1(I) collagen gene in chondrocytes (Fig. 1A),
while sense strand transcription was very low relative to the levels of
antisense transcription and also relative to the levels of sense strand
transcription found in BrdU-treated chondrocytes (Fig. 1B) after normalization with the 27 S rRNA probe.
The antisense transcripts arise from RNA polymerase II transcription,
as their synthesis was inhibited by addition of 2 µg/ml of
-amanitin to the nuclear run-on reactions (data not shown). As
expected, very low levels of the fibroblast-specific 2(I) collagen
transcript were detected in chondrocytes (Fig. 1A),
while high levels of transcription of the type II collagen gene were
found. No antisense transcription was detected for either of these
genes, indicating that antisense transcription is not a general
phenomenon in these chondrocytes.
1(I) collagen gene, the four
probes correspond in location to the DNA sequences mapped in Fig. 3as follows: probe 1 = pBt-Int, probe 2 = pBS-466, probe3 = pBtRS500, and probe4 = pBSCg54-344. Only probes2 and 4 are shown for BrdU-treated chondrocytes (B). The 2(I) collagen probe contains 443 nt of the
fibroblast-specific 5` end sequence, and the type II collagen probe
contains 215 nt of the 3`-untranslated region. M13mp18 and M13mp19
single-stranded DNA was used as a negative control, and a
double-stranded cDNA probe (8 µg/slot) of 27 S ribosomal RNA was
used as a positive control.
1(I) collagen
gene showing the location of probes used in RNase protection
experiments. A, 5` end genomic sequence. B, 3` end
cDNA sequence. The approximate scale is shown underneath each
diagram.
BrdU Treatment of Chondrocytes Causes Loss of Antisense
Transcription and a Simultaneous Acquisition of Sense Strand
Transcription
When nuclear run-on assays were performed with
chondrocytes that had been treated with BrdU (Fig. 1B),
a shift to transcription of the sense strand of the 1(I) collagen
gene was seen, with a corresponding loss (probe4),
or at least a 20-fold down-regulation (probe2), of
transcription of the antisense strand. Therefore, this is a reversal of
the situation found in untreated chondrocytes (Fig. 1A). BrdU treatment also caused stimulation of
transcription of the fibroblast type 2(I) collagen gene, along
with a loss of type II collagen transcription relative to untreated
chondrocytes (Fig. 1A). In addition, the BrdU-treated
chondrocytes, like the untreated chondrocytes, showed no antisense
transcription of either the 2(I) or type II collagen genes. This
pattern of transcription is very similar to that given by chick embryo
fibroblasts (data not shown).Antisense Transcripts Are Accumulated at Steady State
Levels in Chondrocytes, But Not in Fibroblast-like Cells
RNase
protection analyses were carried out using strand-specific riboprobes
of the 1(I) collagen gene. In Fig. 2A, a 467-nt
sense strand riboprobe spanning the intron 1-exon 2 junction (the
region mapped as pBS-466 in Fig. 3A) was used. In this
experiment, protected pieces corresponding to 467 nt and also a doublet
of 235-240 nt were observed in chondrocytes, indicating that
endogenous antisense RNA is accumulated at moderate levels in
chondrocytes. This antisense RNA was found only at very low levels in
BrdU-treated chondrocytes and was not detected in chick embryo
fibroblasts or calvaria, consistent with the nuclear run-on data. In
addition, antisense RNA has been detected in chick embryo sternal
chondrocytes that were not grown in culture (data not shown),
indicating that antisense transcription is not an artifact caused by
culture of the chondrocytes. Preliminary data (not shown) also indicate
the presence of antisense transcripts in chick embryo vertebral
chondrocytes and sternal chondrocytes from 18-day-old embryos, but
their relative abundance remains to be quantitated. For all of the
RNase protection experiments, hybridization was performed at stringent
conditions (68 °C in 80% formamide) to eliminate the possibility of
cross-hybridization with other RNA species, and the hybrids could still
be detected at increasing concentrations of formamide, up to 90% (data
not shown).
1(I) collagen gene shown below the autoradiograms were used. A, riboprobe of
sense sequence showing the antisense RNA in the indicated cell types. 5
µg of total RNAs were used, with 10 µg of tRNA as a negative
control. MspI-digested pBR322 DNA fragments were used as size
markers, although the RNA fragments migrate more slowly and are
therefore smaller than indicated by comparison with these size markers. B, riboprobe of antisense sequence showing the sense mRNA,
with 3 µg of total RNAs (same preparations as in A) and 10
µg of tRNA. The positions of the spliced and unspliced bands are
indicated. CEF, chick embryo
fibroblasts.
1(I) collagen mRNA were detected in
chondrocytes, while high levels were seen in BrdU-treated chondrocytes
and chick embryo fibroblasts, with maximal levels seen in calvaria,
where some unspliced mRNA and splicing intermediates were also
detected. In addition, this experiment shows that the low level of mRNA
found in chondrocytes is correctly spliced, as only 195 nt
corresponding to the exon 2 portion of the probe is protected. It is
also known that this low level of mRNA is polyadenylated, is
transported to the cytoplasm, and is initiated at the normal exon 1
site (data not shown). It is possible that the presence of this mRNA is
due to fibroblast contamination of the chondrocyte cultures, which may
also be the source of the low levels of sense strand transcription
found in the nuclear run-on assays.1(I) collagen gene
mapped in Fig. 3and indicated in Table 1. These probes
spanned various regions of the gene, including one starting at position
-221 of the promoter to +93 of exon 1 (pBt-314), three
others spanning from the HinfI site in intron 1 to the EcoRI site 83 nt into exon 5 (pBt-Int, pBS-466 and pBtRS500; Fig. 3A) and a 3` end probe containing 345 nt of the
carboxyl-terminal cDNA sequence starting 21 nt after the triple helical
coding region (pBSCg54-344; Fig. 3B). In each
case, when a sense sequence riboprobe was used, antisense RNA was
detected at moderate levels in chondrocytes, but not in BrdU-treated
chondrocytes (Table 1). In all cases, as in Fig. 2A, multiple sized pieces of the probes were
protected, usually a piece corresponding to the full size of the
1(I) collagen portion of the probe, as well as smaller discretely
sized species which, when added together, corresponded to the full size
of the probe. This phenomenon is believed to be due to mismatches
between the sequences of the probes and the RNAs in the present chick
cells, since the probes derive from the genomic clones pRS4.6 and
pRS500, isolated by Finer et al.(15) , and may be
coming from a different strain of chicks. In the case of all the 5` end
probes, these putative mismatches have been mapped to intronic regions,
where polymorphisms are more likely to be found.
1(I) collagen mRNA was
detected at substantial levels only in the fibroblast-like BrdU-treated
chondrocytes (Table 1). In all of these RNase protection
experiments, the antisense RNA was present in higher abundance in
chondrocytes than the low level of sense 1(I) collagen mRNA;
however, the antisense RNA in chondrocytes was accumulated at a lower
level than the sense mRNA in fibroblast-like cells.1(I) collagen locus. In
addition, sense and antisense riboprobes from the 3`-untranslated
region of the type II collagen gene (pBSCg12-215) showed that
type II mRNA was accumulated at a high level only in chondrocytes,
while no type II antisense RNA was found in any cell type (Table 1). Again, in agreement with the nuclear run-on assays,
this indicates that the antisense transcription is specific for the
1(I) collagen gene.The
Control experiments were carried out in order to
demonstrate that the antisense transcripts accumulated in chondrocytes
are authentic RNAs and are not some artifactual or contaminating DNA
species which hybridize to the riboprobes. To show this, RNase
protection reactions were performed as in Fig. 2A, but
the same sense sequence riboprobe was hybridized to two samples of
chondrocyte total RNA, as well as to two samples of the pBS-466
riboprobe template DNA prior to the usual RNase digestion step. After
this digestion and before loading the resulting RNase protected hybrids
on the gel, one chondrocyte and one template DNA sample were treated
with the enzyme RNase V1, which is specific for double-stranded RNAs,
but will not act on an RNA-DNA hybrid. In Fig. 4A, it
can be seen that the treated chondrocyte sample was susceptible to
RNase V1 digestion, clearly indicating that the nucleic acid hybridized
to the riboprobe is RNA. As expected, the riboprobe-DNA hybrid was not
susceptible to RNase V1 digestion. In Fig. 4B, a
similar and complementary experiment was carried out, this time using
the enzyme RNase H, which digests the RNA strand of an RNA-DNA hybrid.
In the case of the treated chondrocyte sample, the riboprobe was not
susceptible to RNase H digestion, indicating that the other strand of
the hybrid is not DNA, so therefore must be RNA, consistent with the
results in Fig. 4A. On the other hand, the riboprobe
was digested by RNase H when it was hybridized to DNA, indicating that
the enzyme is active.1(I) Collagen Antisense Transcripts Are
Authentic RNAs
1(I)
collagen gene shown below the autoradiograms was used. The
riboprobe was hybridized to 8 µg of chondrocyte total RNA, 0.2
µg of pBS-466 riboprobe template DNA, or 10 µg of tRNA (t) as indicated above each lane. The RNase
protected hybrids were treated (+ lanes) with either
RNase V1 (A) or RNase H (B). The - lanes indicate samples that have not been treated with these enzymes.
Marker (M) and probe (P) lanes are also
shown. The smear in the - lane with pBS-466 is likely to
come from the competition of the riboprobe with the denatured pBS-466
sense strand for hybridization with the pBS-466 template (antisense)
strand, with the result that varying lengths of the riboprobe
hybridized to the template strand and so survived the digestion with
RNases A and T1.
1(I) collagen antisense sequences. This conclusion
has also been demonstrated by strand-specific RNA-polymerase chain
reaction, where a 175-base pair partial antisense cDNA from exon 2 was
generated in chondrocytes, and subsequent sequencing of this product
revealed that it is identical to the corresponding region of the
1(I) collagen gene (data not shown).The Antisense Transcripts Are Large and Heterogeneous in
Size
In order to find out the sizes of the antisense
transcripts, Northern blots were performed using both 5` and 3` end
strand-specific probes in M13 vectors (Fig. 5). The 5` end
probes contained 467 nt of the intron 1-exon 2 junction, while the 3`
end probes contained 345 nt of the carboxyl-terminal sequence of the
1(I) collagen gene, as mapped in Fig. 3, A and B, respectively. In each case, when a sense sequence probe was
used (Fig. 5A), hybridization to multiple antisense
transcripts was seen in chondrocyte RNA but not in BrdU-treated
chondrocyte RNA. At least two transcripts of >10 kb and larger than
the 45 S precursor rRNA were observed in chondrocytes, and the same
transcripts were seen for both the 5` and 3` end probes, indicating
that both of these regions are included in these transcripts. This
suggests that continuous antisense transcription is occurring across
the 1(I) collagen gene in chondrocytes, including the intronic
regions. Presumably, these transcripts originate at a more downstream
region of the 1(I) collagen gene that has yet to be identified.
1(I) collagen gene were
used. A, antisense transcripts in chondrocytes (C
lanes) or BrdU-treated chondrocytes (B lanes), with 10
µg of total RNA per lane. B, sense strand transcripts,
with 5 µg of total RNA per lane. The positions of the
ribosomal RNA markers are shown in each
case.
1(I) collagen mRNA was
seen at appreciable levels in BrdU-treated chondrocytes and at very low
levels in untreated chondrocytes. Here, the mRNA species seen are the
characteristic 4.7- and 4.9-kb doublet, presumed to be due to two
polyadenylation signals, in addition to the larger and less abundant
6.7-kb uncharacterized species also detected by previous
workers(3, 6, 23, 24, 25) .
The very large species of >10 kb were not detected with these probes
in either cell type, even on long exposures of the autoradiograms,
indicating that these are antisense transcripts and are not due to
unprocessed 1(I) collagen transcripts, as previously
suspected(3) .The Antisense Transcripts Are Moderately
Stable
The stability of the antisense transcripts in
chondrocytes was measured by inhibition of transcription with
actinomycin D, and RNA was extracted at various time intervals in order
to examine its decay rate by RNase protection analysis. In Fig. 6A, a 467-nt intron 1-exon 2 sense sequence
riboprobe of the 1(I) collagen gene, corresponding to the region
mapped as pBS-466 in Fig. 3A, was used to examine the
antisense RNA decay rate. Disappearance of this RNA can be seen from
4 h after actinomycin D treatment, and subsequent densitometry
scanning and analysis revealed that it has a half-life of
6-7 h. Therefore, this antisense RNA is more short-lived
than its corresponding sense mRNA in fibroblast-like cells, previously
shown to have a half-life of 12 h(8) . As a control to
show the integrity of the RNA samples from the actinomycin D-treated
chondrocytes, the decay rate of type II collagen mRNA was also examined (Fig. 6B). In contrast to the 1(I) collagen
antisense RNA, this mRNA is very stable with a half-life of 15 h,
as found in the previous study by Askew et al.(8) . In
addition, 27 S rRNA is shown to be extremely stable in these samples (Fig. 6C), as expected.
1(I) collagen antisense RNA. A 467-nt riboprobe spanning the
intron 1-exon 2 junction was used with 10 µg of total RNA from
actinomycin D-treated chondrocytes at each time point. The various time
points are indicated above each lane, as are probe (P) and marker (M) lanes. B, decay of type
II collagen mRNA. A 215-nt riboprobe from the 3`-untranslated region
was used with 2 µg of total RNA from each time point. C,
decay of 27 S ribosomal RNA. A 180-nt riboprobe was used with 1
µg of total RNA from each time point.
The Antisense Transcripts Are Apparently
Chondrocytespecific
In addition to chondrocytes, BrdU-treated
chondrocytes and chick embryo fibroblasts, a variety of other tissues,
as indicated in Fig. 7, were also examined for the presence of
1(I) collagen antisense RNA at the steady state level. A riboprobe
spanning the intron 1-exon 2 region of the 1(I) collagen gene
shows that antisense RNA is accumulated in chondrocytes only and is not
found in any of the other tissues tested. In addition, these tissues do
not accumulate 1(I) collagen mRNA to substantial levels either, as
compared with fibroblast-like cells (data not shown). As a positive
control, the presence of RNA in each sample was examined using a
180-nt 27 S rRNA riboprobe, as shown under each lane. Therefore,
insofar as has been tested, the accumulation of 1(I) collagen
antisense RNA is chondrocyte-specific.
1(I) collagen antisense RNA in a variety of
tissues. A 351-nt riboprobe from the region shown below the
autoradiograms was used, with 6 µg of total RNA from each tissue,
and 10 µg of tRNA as a negative control. Marker (M) and
probe (P) lanes are also indicated. In addition, the
presence of 27 S ribosomal RNA is shown in each sample, using a
180-nt riboprobe and 2 µg of total RNA per
lane.
The
In order to determine if the
antisense transcripts were derived from the same locus as the 1(I) Collagen Gene Is Apparently Present in a
Single Copy in the Chick Genome1(I)
collagen mRNA, genomic Southern blots were performed to see if the
1(I) collagen gene is present in a single copy in the chick
genome. Since it has previously been shown that two types of repetitive
sequences, also found at other loci in the chick genome, are located in
the promoter region and the 5`-most part of intron 1 of the 1(I)
collagen gene (15) , care was taken to avoid these regions in
the Southern blot analysis. Therefore, a DNA sequence spanning the end
of intron 1 to the end of exon 2, corresponding to pBS-466 in Fig. 3A, and not known to contain repetitive sequences,
was used to probe genomic DNA that had been digested separately with
four different restriction enzymes. This sequence was cloned in an M13
vector (466-mp18), since single-stranded probes were necessary for the
chemiluminescent detection kit used for these Southern blots. In Fig. 8, it can be seen that single bands were found in each
lane, where BamHI digestion gives a 0.63 kb band, and StuI digestion gives a 5.1 kb band, as predicted from the
restriction map of the 1(I) collagen 5.1-kb genomic clone
SA/S51(15) , with EcoRI and HindIII
digestions giving single bands of >5.1 kb, since these sites are not
present in the genomic clone. In addition, very high molecular weight
bands were also seen in each lane near the top of the gel (not shown),
but control experiments showed that these were due to hybridization by
the M13 vector alone and were not specific for the 1(I) collagen
gene (data not shown). In addition, single 1(I) collagen gene
bands, along with the same high molecular weight vector bands, were
also detected using a single-stranded cDNA probe (344-mp18,
corresponding to pBSCg54-344 in Fig. 3B),
containing part of the carboxyl-terminal coding sequence (data not
shown). Therefore, the 1(I) collagen gene appears to be in a
single copy in the chick genome (although the presence of an exact
duplicate cannot be ruled out by this analysis), suggesting that the
antisense transcription is also coming from this locus.
Transcription and Accumulation of Antisense RNA in
Chondrocytes Simultaneously with the Down-regulation of
In this report it is shown that antisense
transcription of the 1(I)
Collagen mRNA1(I) collagen gene is occurring at a high
rate in chick embryo sternal chondrocytes. This explains the high rate
of transcription found in the previous study when double-stranded
probes were used(8) . It is also shown that the antisense
transcripts are accumulated at moderate levels in chondrocytes, and
probes from various locations show that antisense transcription is
occurring throughout the gene, including the intronic regions and at
least through some of the 1(I) collagen promoter region. The
accumulation of these transcripts appears to be specific for
chondrocytes, as they have not been detected in any other tissues
tested, although the possibility remains that they could be transcribed
in these tissues and then rapidly degraded, or they could be
transcribed in other tissues yet to be tested and/or at other stages in
development.1(I) collagen mRNA (sense strand)
transcription. This finding is in contrast to previous work (3, 8) which suggested that sense strand transcription
was occurring rapidly in chondrocytes, and it also provides an
explanation for the very low steady state levels of 1(I) collagen
mRNA in these cells. However, it is still possible that some
post-transcriptional down-regulation occurs so rapidly that significant
levels of transcription escape detection by the nuclear run-on
procedure. It is unclear if the low level of mRNA detected in
chondrocytes is due to a low level of fibroblast contamination of the
chondrocyte cultures, as previously suspected(8) , or whether
its presence is due to authentic chondrocyte transcription occurring at
a low level.Loss of Antisense Transcription by BrdU Treatment of
Chondrocytes
Treatment of chondrocytes with BrdU causes an
induction of sense strand transcription and, concurrent with this, a
loss, or at least a great reduction, of antisense transcription.
Therefore, an antiparallel relationship exists between chondrocytes and
BrdU-treated chondrocytes with regard to which strand of the 1(I)
collagen gene is being transcribed. It appears that transcription of
only one strand can be tolerated in these cells at a time, where
transcription of the complementary strand is down-regulated, perhaps
due to a less active promoter and/or interference by a more active
transcription complex coming in the opposite direction, although
extremely rapid post-transcriptional degradation cannot be ruled out.
Since BrdU-treated chondrocytes resemble undifferentiated mesenchymal
cells which produce type I collagen(26) , the turn-off of
antisense transcription during BrdU treatment suggests that the
antisense transcripts are not present in mesenchymal cells but are
acquired during the reverse process of chondrogenesis.Size and Stability of the Antisense
Transcripts
The large sizes of the antisense transcripts are
consistent with the notion that antisense transcription is occurring
across the entire 1(I) collagen gene in chondrocytes, as the same
transcripts were detected using both 5` and 3` end probes. The exact
size of the chick 1(I) collagen gene is not known, but the human
gene is known to be 18 kb long(27) . In higher eukaryotes this
gene has a conserved pattern of introns and exons, where differences in
intron sizes are responsible for different gene lengths(27) .
If there is conservation in gene size between human and chick, then the
antisense transcripts are long enough to extend all across the chick
gene. Since the ends of the antisense transcripts have not been
identified, it is not known how far upstream or downstream of the
1(I) collagen gene these transcripts extend. Presumably, the
antisense transcripts have their own chondrocyte-specific promoter that
initiates formation of an RNA polymerase II transcription complex,
located in a more 3` or downstream region of the 1(I) collagen
gene. It is not known why multiple transcripts are detected, but this
could possibly be due to multiple start or termination sites,
differential processing of the transcripts, or it may be an
experimental artifact due to the very large sizes of the transcripts,
where secondary structures and partial renaturation lead to
differential separation in the gel.1(I) collagen
mRNAs is not correct, and, instead, these are antisense transcripts. In
addition, transcripts of 10 kb in chondrocytes from 16-day-old
embryos have been reported by another group(6, 23) .
The failure to detect these transcripts in other studies may be
reflective of their lower abundance and stability compared to the sense
mRNA in fibroblast-like cells. In the present study, the antisense
transcripts in chondrocytes were not as readily detectable as the sense
transcripts in fibroblast-like cells, especially by Northern blot
analysis; however, in both RNase protection assays and Northern blots
the antisense transcripts were easier to detect than the low level of
sense transcripts in chondrocytes.1(I) collagen gene in chondrocytes (see
below). The existence of relatively large and seemingly functional
untranslated RNAs have been described in eukaryotes before; for
example, the 17-kb human XIST transcript which is confined to the
nucleus and is believed to be involved in the inactivation of the X
chromosome(30) , the cytoplasmic H19 mouse and human transcript
which has tumor-suppressor activity(31, 32) , and the
developmentally regulated dut A RNA of Dictyostelium
discoideum(33) .1(I) sense mRNA in
BrdU-treated chondrocytes. This is shown here, where the half-life of
the antisense transcripts is 6-7 h, compared with 15 h
for type II collagen mRNA, or 12 h for the sense 1(I)
collagen mRNA(8) . If indeed antisense transcription plays a
regulatory role in expression of the 1(I) collagen mRNA in
chondrocytes, then these transcripts need not be very stable.Naturally Occurring Antisense RNA in
Eukaryotes
Increasing numbers of naturally occurring antisense
RNAs are being found in eukaryotes, although in most cases their
function is as yet unknown (reviewed in (34, 35, 36, 37) ). Relatively few
cell or stage type-specific antisense RNAs have been described. One
such example exists in C. elegans, where the short
(22 and 40 nt) lin-4 transcripts have complementarity to a
region repeated seven times in the 3`-untranslated region of the
developmentally regulated lin-14 transcript, and base pairing
of these complementary regions can inhibit translation of the lin-14 mRNA at the appropriate developmental
stage(38, 39) . In this case, as in many of the other
endogenous antisense RNA cases described, the region of complementarity
is not complete, or else there is only partial overlap of the
transcripts. A stage type-specific antisense RNA also regulates the Dictyostelium EB4-PSV pre-spore gene, apparently at the level
of mRNA stability(40) . In this case, the EB4-PSV mRNA is
transcribed constitutively during development but becomes unstable in
vegetative and disaggregating cells, when the antisense RNA, which
lacks coding capacity, is present. During the induced differentiation
of murine erythroleukemia cells, an induced nuclear RNA (inRNA) that is
complementary to the first intron of the p53 gene is accumulated, and
may regulate processing of the p53 mRNA (41) .Possible Roles of the
The 1(I) Collagen Antisense
Transcripts1(I) collagen antisense transcripts
represent a new case of cell type-specific antisense RNA. Unlike most
cases in which antisense transcripts have been found, they exist in
chondrocytes concurrent with little or no sense strand transcription.
These antisense transcripts are also unusual in that they are very
large and appear to extend across the entire 1(I) collagen gene,
and the complementarity with the regions tested is very strong, if not
complete, as high stringencies were used in the RNase protection
experiments. Since these antisense transcripts are present in a
differentiated and specialized cell type, it poses the question of why
chondrocytes should acquire such very large transcripts. Their cell
type-specific presence suggests that they have a functional role,
although the possibility that they are evolutionary artifacts or
accidental read-through products of an adjacent gene cannot be ruled
out. Obviously, a demonstration of these transcripts in chondrocytes
from other species would reinforce the suggestion that they play a
functional role.1(I) collagen gene in
chondrocytes, especially since the induced turn-on of sense strand
transcription, by BrdU treatment of chondrocytes, is accompanied by the
disappearance of these antisense transcripts. Since both transcripts
appear to derive from the same piece of DNA, it is possible that
collision of two transcription complexes coming in opposite directions
occurs, resulting in premature termination or a reduced elongation rate
of one transcript, while the transcript with the highest transcription
rate predominates. In chondrocytes this would be the antisense
transcript, presumably because of a highly active chondrocyte-specific
promoter. In BrdU-treated chondrocytes, on the other hand, a strong
turn-on of the 1(I) collagen promoter, and perhaps a
down-regulation of the antisense promoter, would lead to predominance
of sense strand transcription.1(I) collagen mRNA transcription, by preventing
binding of important transcription factors in this region.
Transcriptional interference in promoter regions, otherwise known as
promoter occlusion, has been described previously (42, 43, 44) . However, if this was the sole
purpose of the antisense transcripts in chondrocytes, then it is
unclear why these transcripts would span the entire gene, originating
several kilobases away. It is therefore possible that antisense
transcription serves to interfere with both initiation and elongation
of transcription of the 1(I) collagen mRNA.1(I) collagen
transcripts in chondrocytes, that these transcripts could form in
vivo duplexes with the antisense transcripts, which are then
subjected to rapid degradation by unwindases or RNases specific for
double-stranded RNAs. This is the mechanism by which many other
antisense RNAs are presumed to work when they coexist in the same cell
as sense strand transcripts(35, 36) . If a duplex
formed and was degraded very rapidly, then only the RNA that is most
abundant would be detected; in the case of the 1(I) collagen gene,
the antisense RNA in chondrocytes.1(I) collagen gene, then
apparently the mechanism would be unique to chondrocytes, as most of
the other tissues tested also contain a down-regulated 1(I)
collagen gene, yet they do not accumulate these antisense transcripts.
It is possible that transcription of the mRNA is not sufficiently
down-regulated at the level of promoter activity in chondrocytes, so an
additional mechanism is needed. The intrinsic activity of the 1(I)
collagen promoter in chondrocytes, in the absence of antisense
transcription, is not yet known.1(I) collagen gene, by transcriptional
interference or by rapid post-transcriptional events, or whether they
encode a chondrocyte-specific product, their unusual presence merits
further investigation, which may provide insight into the function of
naturally occurring antisense RNAs in eukaryotes.
)
We thank Louis Gerstenfeld for provision of the
plasmids pRS4.6 and pRS500 and Linda Kosturko for use of her
densitometer.
©1995 by The American Society for Biochemistry and Molecular Biology, Inc.
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  X. Zhou, R. Sunkar, H. Jin, J.-K. Zhu, and W. Zhang Genome-wide identification and analysis of small RNAs originated from natural antisense transcripts in Oryza sativa Genome Res., January 1, 2009; 19(1): 70 - 78. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. R. Henz, J. S. Cumbie, K. D. Kasschau, J. U. Lohmann, J. C. Carrington, D. Weigel, and M. Schmid Distinct Expression Patterns of Natural Antisense Transcripts in Arabidopsis Plant Physiology, July 1, 2007; 144(3): 1247 - 1255. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. M. Jaworski, M. Beem-Miller, G. Lluri, and R. Barrantes-Reynolds Potential regulatory relationship between the nested gene DDC8 and its host gene tissue inhibitor of metalloproteinase-2 Physiol Genomics, January 17, 2007; 28(2): 168 - 178. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Haddad, A. X. Qin, P. W. Bodell, L. Y. Zhang, H. Guo, J. M. Giger, and K. M. Baldwin Regulation of antisense RNA expression during cardiac MHC gene switching in response to pressure overload Am J Physiol Heart Circ Physiol, June 1, 2006; 290(6): H2351 - H2361. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. E. Coudert, L. Pibouin, B. Vi-Fane, B. L. Thomas, M. Macdougall, A. Choudhury, B. Robert, P. T. Sharpe, A. Berdal, and F. Lezot Expression and regulation of the Msx1 natural antisense transcript during development Nucleic Acids Res., September 12, 2005; 33(16): 5208 - 5218. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Chao and A. P. Spicer Natural Antisense mRNAs to Hyaluronan Synthase 2 Inhibit Hyaluronan Biosynthesis and Cell Proliferation J. Biol. Chem., July 29, 2005; 280(30): 27513 - 27522. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. G. Elmendorf, S. M. Singer, and T. E. Nash The abundance of sterile transcripts in Giardia lamblia Nucleic Acids Res., November 15, 2001; 29(22): 4674 - 4683. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. A. Thrash-Bingham and K. D. Tartof aHIF: a Natural Antisense Transcript Overexpressed in Human Renal Cancer and During Hypoxia J Natl Cancer Inst, January 20, 1999; 91(2): 143a - 151a. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. B. Bahramian and H. Zarbl Transcriptional and Posttranscriptional Silencing of Rodent alpha 1(I) Collagen by a Homologous Transcriptionally Self-Silenced Transgene Mol. Cell. Biol., January 1, 1999; 19(1): 274 - 283. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Borgonovo, G. Racchetti, M. Malosio, R. Benfante, P. Podini, P. Rosa, and J. Meldolesi Neurosecretion Competence, an Independently Regulated Trait of the Neurosecretory Cell Phenotype J. Biol. Chem., December 25, 1998; 273(52): 34683 - 34686. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Kumar and G. G. Carmichael Antisense RNA: Function and Fate of Duplex RNA in Cells of Higher Eukaryotes Microbiol. Mol. Biol. Rev., December 1, 1998; 62(4): 1415 - 1434. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Hartmann, A. Krug, M. Bidlingmaier, U. Hacker, A. Eigler, R. Albrecht, C. J. Strasburger, and S. Endres Spontaneous and Cationic Lipid-Mediated Uptake of Antisense Oligonucleotides in Human Monocytes and Lymphocytes J. Pharmacol. Exp. Ther., May 1, 1998; 285(2): 920 - 928. [Abstract] [Full Text]
|
|
|
|
|
|
G. Ambrosini, C. Adida, G. Sirugo, and D. C. Altieri Induction of Apoptosis and Inhibition of Cell Proliferation by survivin Gene Targeting J. Biol. Chem., May 1, 1998; 273(18): 11177 - 11182. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Ansaldi, A. Chaboud, and C. Dumas Multiple S Gene Family Members Including Natural Antisense Transcripts Are Differentially Expressed during Development of Maize Flowers J. Biol. Chem., July 28, 2000; 275(31): 24146 - 24155. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| All ASBMB Journals | Molecular and Cellular Proteomics |
| Journal of Lipid Research | ASBMB Today |